Deploying a new cluster

This section provides instructions for deploying a new NDR multibox cluster. If you need to expand your existing NDR cluster, see Expanding an existing NDR cluster deployment.

The deployment process entails creating configuration files for each node and then copying the files to the appropriate device. You can configure an NDR cluster through the cluster setup and maintenance menu in KLISH.

Note
During the NDR first run on the CLI, you are required to select if the NDR is part of a multi-box cluster or a single-box cluster. Select 'y' for 'yes' if it is a multi-box cluster otherwise select 'n' for 'no'.

This table summarizes the node configuration.

Node Configuration

Sample Setting

(N) Node name:

ia1

Must be unique within the cluster.

Elasticsearch does not allow uppercase characters in index names. Cluster names, used to create index names, and hostnames must be lowercase when entered in KLISH.

(M) Device is configured as master node:

True

(D) Device is configured as data node:

True

(A) Disable _all field in ES Index:

False

Create Master Node Configuration

Before you create the Cluster Master node, you must set up the initial NDR configuration.

Configure the cluster name with a unique name.
Configure the node name with a unique name. This name is the NDR hostname.
Configure the network settings for eth0 and eth1.
Apply the configuration settings, and then reboot into the NDR appliance.
Ensure that the interfaces have the configured settings.

This table summarizes the cluster communication settings.

Cluster Communication	Sample Settings
(R) Cluster name:	clus-1 This name must be the same for all nodes in the cluster. Caution Elasticsearch does not allow uppercase characters in index names. Cluster names, which are used to create index names, and hostnames must be lowercase when entered in KLISH.
(U) Unicast ping IP address (port):	192.168.1.1:9300 This is a private `eth1` IP address of the master. In a NDR mulltibox cluster, the data node Unicast ping address should match the `eth1` IP address of the master data node and default port of 9300.
(B) Bind host IP address:	192.168.1.1 The IP addresses for cluster nodes discovery requests, always set to the private `eth1` IP address of the NDR data node that you are configuring.
(P) Publish host IP address (port)	192.168.1.1 Always set to the private `eth1` address of the NDR data node that you are configuring.
(S) Per Index default shard count:	4 Set the shard count to 4 times the number of NDR appliances in your multi-box NDR cluster. For example, if you have one NDR host in a cluster, set the shard count to 4. If you have two NDR hosts in a cluster, set the shard count to 8.
(H) HTTP enabled:	True
(T) Multi-NDR Cluster:	Yes
(G) Gateway Recovery node count:	This option must be configured with 2 for a single node cluster and with 4 for a multinode cluster.

For more information, see Initial configuration.

To start the cluster setup:

Log in to the CLI on the NDR.
Enter privileged mode:
hostname > enable
Enter the admin password:
[sudo] password for <npadmin>: <password>
Enter configuration mode:
hostname # configure system
Enter cluster setup:
user@hostname(config)# cluster setup
Note
The script indicates that its purpose is for configuring NDR Clusters and not for Distributed Search (director).
Choose yes to continue.
From the NDR Cluster Configuration menu, choose
(E) Edit/review cluster configuration.
From the edit configuration menu, choose the following:
- Node name: Must be unique within the cluster.
- Master node: true
- Data node: true
- Cluster name: Must be lowercase and must be the same for all nodes in the cluster. Note this name for data node setup.
- Unicast ping address: Change the Unicast ping IP address and add <eth1-ip>:9300 (e.g. 192.168.2.1:9300). The data node Unicast ping NDR address should match the .eth1 IP address of the master data node and default port of 9300.
- Bind host IP address: Specify the IP address(es) that the master should listen on. Set this to the private eth1 IP of the NDR appliance you are configuring.
- Publish host IP address (port): Specify the IP address the to accept connection requests. Set this to the private eth1 IP of the NDR appliance you are configuring.
- Per index default shard count: Set the shard count to 4 times the number of NDR appliances in your multi-box NDR cluster. For example, if you have one NDR host in a cluster, set the shard count to 4. If you have two NDR hosts in a cluster, set the shard count to 8.
- HTTP enabled: true
Your configuration should look similar to this:
```
Editing NDR Cluster Configuration
```
```
------------------------------------------------------------------------------
```
```
Node Configuration
```
```
(N) Node name:                            ia1
```
```
(M) Device is configured as master node:  true
```
```
(D) Device is configured as data node:    true
```
```
Cluster Communication
```
```
(R) Cluster name:                         clus-1
```
```
(U) Unicast ping IP address (port):       192.168.1.1:9300
```
```
(B) Bind host IP address:                 192.168.1.1
```
```
(P) Publish host IP address (port):       192.168.1.1
```
```
(S) Per index default shard count:        4
```
```
(H) HTTP enabled:                         true
```
```
-----------------------------------------------------------------------
```
```
Q: Quit and return to main menu
```
```
Select an option:
```
Select T to turn on the cluster.
Press the Q key to quit and return to the main menu.

To save and transfer the cluster configuration:

Select S to save and transfer cluster configuration.
When prompted to enter the configuration file name, accept the default or enter the desired name.
The screen indicates the location of the configuration file.
Choose an option:

To install the configuration on the local machine, indicate yes when prompted.
To transfer the configuration to a remote host to be installed later, indicate yes when prompted.

The system will prompt for the remote host name. The npscp account will need to be enabled on the remote host, and you will need the credentials for that account.
When prompted for a destination location, accept the default.

Create data node configuration

You can configure each data node using the CLI cluster setup menu.

To create a data node configuration file:

On the NDR, log in to the CLI.
Enter privileged mode:
hostname > enable
Enter the admin password:
[sudo] password for <npadmin>: <password>
Enter configuration mode:
hostname # configure system
Enter cluster setup:
user@hostname(config)# cluster setup
Note
The script indicates that its purpose is for configuring NDR clusters and not for distributed search (director).
Choose yes to continue.
From the NDR cluster configuration menu, choose
(E) Edit/review cluster configuration.

From the edit configuration menu, choose the following:

Node Configuration	Setting
Node Name	Must be unique within the cluster.
Master Node	False.
Data Node	True.
Cluster Name	Must match the cluster name specified in the master node configuration.
Unicast ping address	Must match the private `eth1` IP master node IP address and default port of 9300. For example, 192.168.1.1:9300.
Bind IP address	Specify any IP address that the data node should listen on. This address should match the address for the `eth1` interface. For example, 192.168.1.2.
Publish host IP address	Specify any IP address that the data node should publish on. This address should match the address for the `eth1` interface. For example, 192.168.1.2.
Per index default shard count	Set the shard count per indexed data to 4 times the number of NDR hosts in a your NDR cluster. For example, If you have one NDR host in the cluster, set the shard count to 4. If you have two NDR hosts in the cluster, set the shard count to 8.
HTTP Enabled	False.

Select T to turn on the cluster.
Press the Q key to quit to the main menu.
When prompted to enter the configuration file name, accept the default or enter the desired file name.
The screen indicates the configuration file location.
Choose an option:

To install the configuration to the local machine, indicate yes when prompted.
To transfer this configuration to a remote host to be installed later, indicate yes when prompted.

The system prompts for the remote host name. The npscp account must be enabled on the remote host. You will need the credentials for that account.

Disable unnecessary services on data node

Several services that are enabled from the default NDR configuration are unnecessary for Data Node operation. Disable these unnecessary services.

To disable unnecessary services on data node:

Log into the CLI on the NDR.
Enter privileged mode:
hostname > enable
Enter the admin password:
[sudo] password for <npadmin>: <password>
Enter configuration mode:
hostname # configure system
Enter cluster hardening:
user@hostname(config)# cluster hardening enable
When services are disabled, you should see this message:
Services disabled and stopped.

Start the master node

Your cluster is now properly configured. Reboot the NDR appliance.

To reboot the master node:

Log in to the CLI on the NDR.
Enter privileged mode:
hostname > enable
Enter the admin password:
[sudo] password for <npadmin>: <password>
Reboot the appliance.
hostname # reboot

Start the Data Node

After the master node is running, you can restart each data node.

To reboot the data node:

Log in to the CLI on theNDR.
Enter privileged mode:
hostname > enable
Enter the admin password:
[sudo] password for <npadmin>: <password>
Reboot the appliance.
hostname # reboot

Verify the Cluster State using the CLI

When at least one data node is connected, you can query the master node with the show cluster command. This command displays Cluster Health and Node. The cluster health section gives a high-level summary of the cluster status. An important indicator here is status, which should be green.

The node status section indicates which nodes are connected to the cluster and which role they play. You will see the master and data nodes that are configured in this NDR Cluster.

To verify the cluster on the master node:

Log into the CLI on the NDR.
Enter privileged mode:
hostname > enable
Enter the admin password:
[sudo] password for <npadmin>: <password>
Display the cluster:
user@hostname(config)# show cluster

Verify the cluster state using the API

You can verify the cluster state using Elasticsearch web APIs.

To verify the cluster state using the Web UI and API:

Log in to the NDR Web UI. Open a Web browser and type https://appliance in the address line, where appliance is the IP address or hostname of the appliance. For example, if the configured IP address of the appliance is 10.1.0.1, type:
https://10.1.0.1
Enter this URL, where iaipaddress is the IP address to the NDR master node:
https://iaipaddress/elasticsearch/_cat/nodes
Check the overall cluster health by entering this URL where iaipaddress is the IP address to the NDR master node:
https://iaipaddress/elasticsearch/_cat/health